Counting benefits does not much change the income stagnation story

Lane Kenworthy reports:

A third worry is that the income measure used to calculate median family income is too thin. If a growing portion of GDP has gone to employer benefits, that would help middle-class households, but it wouldn’t show up in these income data.

To address these second and third concerns, we can turn to a more encompassing measure of household income. The data are from the Congressional Budget Office (CBO). The measure includes all sources of cash income. It adds in-kind income (employer-paid health insurance premiums, food stamps, Medicare and Medicaid benefits), employee contributions to 401(k) retirement plans, and employer-paid payroll taxes. Tax payments are subtracted.

We can use average household income in these data as a substitute for GDP per capita. The CBO data set doesn’t tell us the median income, but it provides something quite similar: the average income of households in the middle quintile of the distribution (from the 40th percentile to the 60th). The following chart adds these two series. The story is virtually identical.

He considers some other adjustments too, and this is the final story:

Addendum: Matt Yglesias comments.


Doesn't this illustrate the major fallacy of "The Great Stagnation"? Perhaps that's why there is no mention of Tyler's book.

No it doesn't. Also, redistribution is not a safety net. Using words wrong is going to be the downfall of the left.

(That is your links bottom line, that stagnation isn't so bad as long as the winner-take-all winners provide a safety net to the losers. To make a long story short, that's not going to fly.)

"Using words wrong is going to be the downfall of the left."

On the contrary, it's their stock and trade. There's always a new generation that is unwise and inexperienced and ripe to be lead by the newest pied piper.


Actually, the book summarizes one discussion on income inequality with something that I think you might agree with: "The 'rise in income inequality' and the 'slowdown in ideas production' are two ways of describing the same phenomenon, namely that current innovation is more geared to private goods than to public goods."

File this under "Dangers of judging a book by reading a blog post about a talk about the book."

Hello, moderator.

Could this be from larger government spending which has moved the redistribution towards government workers (educated higher wage workers, teachers, people with masters degrees, researchers, etc)?

Furthermore, the breakpoint appears to be in the 70's when the US started the "war on poverty." Could that have something to do with it?

I think so. I suspect that the 40-60 crowd has been trading transfer payments for work -- which is a big net gain in leisure and hence utility.

We have what, 50M on food stamps today?

The "War on Poverty" would date back to the 1960s.

Tyler, what am I missing? Couldn't average house hold income be used as a proxy for median family income rather than GDP?

The footnote confused me: "Average Household Income is set to equal GDP per capita in 1979"

Is this normalization kosher? I mean, if you define two series to be equal valued at 1979 aren't they axiomatically going to seem to diverge at that point?

Yes, but it does still seem pretty clear that one grows faster than the other after that point.

I am not a quantitative economist, but I do remember the 70s, 80s, 90s, and 00s.

The standard of living available today on a middle class income is far superior to previous decades (assuming of course that you have a job).

The thing is, it's not really, and the having a job part is kind of a big deal. We just all have cheap electronics that ease the pain a little. This is not the same as real stuff.

Cheap electronics, unlimited amounts of free and cheap media (even if you exclude illegal pirating), gourmet coffee, high quality raw and prepared food and ubiquitous gourmet coffee (provided by evil companies like Starbucks and Whole foods), affordable clothing that differs little from high-end clothing, affordable global travel (taken for granted by gen-x-ers) etc.

Take food...and parenting. Here is parenting today in a nutshell. I found organic kale at my grocery store and almost started crying.

Could you have found organic kale for any price in the 80s? I know where I lived, iceberg lettuce and may be cabbage were the only options, now we have organics and multiple different varieties of salad and pretty much everything else. Some of it's pricey, sure, but luxury goods aren't just cars and sailboats, and what used to be truly luxury is now, in many cases, just special occasion stuff.

Why would organic kale make you cry?

You think that's bad? Try eating inorganic kale.


200 years ago even the lower classes could afford to eat organic kale.

Definitely stagnation!

"Why would organic kale make you cry?"

For happiness!

I actually got to the point where I was almost crying and then decided to try to cry.

But I couldn't. I have become too calloused.

So I can stuff clever food in my face and watch videos on my iphone.

However, I can't live in a safe neighborhood, send my kids to college, have reliable medical coverage, or have my wife not need to work. Good trade off.

Isn't crime significantly lower than it was in the 60s? Doesn't a far higher percentage of the population go to college than did in the 60s? And wasn't increased women's participation in the workforce considered a net social benefit until like two years ago, when everyone decided to do a 180 and decide it was terrible?

These, unlike Angry Birds, are very real gains.

Neighborhoods are on avg much safer, community colleges are quite affordable, 1960s medicine was vastly inferior, and you can get most of that without working at all.

Safe neighborhoods is a positional good. By and large, what makes a neighborhood safe is that it keeps the poor out and preferably far away. No amount of economical growth will ever solve that problem baring some massive improvement in law enforcement.

Gourmet coffee! Two income families (itself something not needed in the bad coffee era) are struggling with more of their income disappearing into housing, healthcare, and education costs, but at least they have gourmet coffee, and from Starbucks no less. Is that an argument?

The argument is "what is real wealth?"

It's difficult to square the circle in our minds, but Baumol's cost disease is part of it, outsourcing is another part, and my little hobbie horse is that technology is not deflation, and things like Web 2.0 and infinite Taylor Swift downloads are definitely improvements, but they aren't quite wealth.

Also, if TC figures this problem out then in my completely dilettante opinion I'd call that Nobel-worthy. But hey, I think Buffett and Munger should get the Nobel too.

How many Taylor Swifts to make up one Joan Baez? I'd like to see the CPI guys figure that one out.

They have vastly better housing and healthcare.

The education pricing problem is pretty ridiculous, but you can homeschool for free.

Ed/Andrew' -- Value is subjective. That society has elected to produce iPhones instead of... whatever... a flying car... doesn't mean "we don't have real wealth," it simply means that your desires are at odds with the majority of society.

But of course, that's true of everyone. There is no such thing as "the average person." If you tally up the features of the apparent representative majority, you'd get a picture of a human being who doesn't really exist. This is the nature of aggregates. In aggregate, we prefer iPhones to flying cars. But if I put that choice to any *individual*, virtually everyone would choose the flying car. That's the mystery of the invisible hand.

"Value is subjective"

That is what I'm disagreeing with.

That's why I say that if TC figures this out he gets the Nobel just for that.

it isn't just electronics. There is ample evidence that the price of most durable household goods, measured in the number of hours one needs to work to obtain them, has fallen substantially in the last 40-50 years. And of course the functionality and quality have increased dramatically.

Have durable household goods improved in functionality and quality in the last 40-50 years as they did in the 40-50 prior to that?

"We just all have cheap electronics that ease the pain a little. This is not the same as real stuff."

We also have substantially more living space per person (much bigger houses) and highly durable cars that are more efficient, require less maintenance and run for more miles. Those certainly represent 'real stuff'. By any objective standards the middle class in America is richer than it was 30 years ago.

It seems to me that the "stagnating" middle class incomes is nothing more than the wage equalization that results from economic growth abroad and immigration. That we are seeing a major wave of anti-immigration sentiment in the developed countries strikes me as evidence in favor of this theory.

In a global sense, we're seeing lots of economic growth and and overall improvement in our standards of living. But the US middle class is probably going to have to suck it up a bit. Their wages aren't going to be increasing like they were when only a handful of countries were developed "super-powers."

Yes ie. skilled labor per se is no longer scarce ...

There is a small amount of solace in the fact that many of the current 1% who have escaped being affected technology and globalization are going to be hit by it soon (ie. universities, legal profession)

If we were still building road networks, as an example, that couldn't be as easily outsourced.

I'm bracing for the reaction from the ivory towers when the chill winds of reality hit. Those people have no idea. The bellyaching will be epic.

If you did want to blame, isn't "free markets" a better target than immigration?

Even if you had zero immigration blue-collar wages would fall so long as foreign made widgets could be imported for cheap.

One point: immigration isn't governed by free markets. In a free market, there would be no such thing as 'immigrants,' because there would only be owners, licensees and trespassers. Movement across property lines would be governed by contract.

I'm fairly sure the average person would call a Guatemalan who rents an apartment next door an immigrant, irrespective of a lack of government involvement.

Also, under such a pure free market, the importance of inherited property may be a little more obvious, and people may be less likely to attribute all of their wealth to their own hard work and natural abilities. Eh, who am I kidding?

Yeah. That's a real problem all right.

My point was: The fall in American blue-collar wages is less a result of the free movement of people and more a result of the free movement of goods and services.

Maybe it doesn't make a difference in the end, but if the comparison purports to be what it says it is, the solid red line (the "all in" compensation) should always be above the solid black line. The fact that it isn't indicates that he is not comparing like with like.

His notation confuses me. There are three "income" lines in the plot; which one of those are "all in compensation" and which ones aren't? Maybe he shouldn't have called them all "income".

The solid red line is family, the solid black line is household, which are slightly different. Also the solid black line is a true median, whereas the solid red line is the "average of all families in the middle quartile," again slightly different.

The median line seems to be tracking the average of the middle quartile very well.

Wonder if it is valid to conclude that the American middle-class is a very equal ( wage ) cohort?

Maybe this has to do with comparing median (exactly the 50th percentile) vs. the average of the middle quartile (not exactly the 50th percentile). Seems unlikely though.

It's the index value - the starting point is 1 in 1947; so since this shows overall growth (not levels) the solid red line can sometimes be below the solid black line.

How can you tell it is showing growth (i.e., change) and not level?

Because it says so in the footnote.

This includes all sources of cash income, including in-kind income like employer-paid health insurance premiums, food stamps, Medicare and Medicaid benefits, employee contributions to 401(k) retirement plans, and employer-paid payroll taxes.

It appears that it doesn't include things like employer-paid "business meals," company cars, and the like, but I'm not sure. The other sources of in-kind income he lists generally show up on payroll information, but I don't think that company meals and company cars tend to do so. They've certainly become somewhat rarer (business meals and entertainment became less deductible, lower marginal rates mean than people would prefer cash to such in-kind sources, etc.)

I don't understand why this is presented as stagnation. I see rising compensation packages.

This does seem more of a inequality argument than a stagnation argument.

That is what I took away as well.

Contrary to Tyler's claim, I'm interested by the effect in counting benefits. It makes little difference from 1979 to 2000, but then starting in 2000 compensation suddenly decouples from income.

The CBO is essentially using compensation data as is used in the productivity data to calculate unit labor cost. The compensation data include all forms of labor income, like health insurance and vacations, except for stock options. I have never seen an estimate of how much excluding stock options distorts the data, but it must be very significant since it started becoming important around 1980.

Can't we use the best numbers all the time? I mean, great, we have included some other types of compensation (probably not all), but we are still looking at "households". Well guess what, "households" is not a meaningful concept. What we should want to know is the compensation of actual people.

Come now, it's not "employer benefits" that are growing. It's health care premiums. Are we really better off because our employers' health costs are exploding? Have our health outcomes improved a corresponding amount in this period of wage stagnation? I don't see the evidence, and in the lack of evidence, I would prefer median family income, or better still, median wage as the proper metric.

Have our health outcomes improved a corresponding amount in this period of wage stagnation?

The prognosis for someone diagnoses with cancer, hearth disease, stroke etc. is much better now than in 1970.

The prognosis for someone diagnoses with cancer, hearth disease, stroke etc. is much better now than in 1970.

Sure, but the quality of my television is much better now than whatever was being produced in 1970, and yet cheaper in real terms. I'm convinced that we could still have realized these health care gains without the cost explosion in a more rational (i.e. market-driven) health care system.

Television in 1972: M.A.S.H.

Television in 2012: M.A.S.H. reruns

Theater in 1612: The Tempest

Theater in 2012: The Tempest

TGS has been going on for longer than I thought!

In the arts, it probably has. I get your joke, but it is a conundrum that Western literature - world literature really - has not produced a writer who surpasses Shakespeare in four centuries.

"It would be to waste the time of my readers and my own if I strove to demonstrate how the general mediocrity of fortunes, the absence of superfluous wealth, the universal desire of comfort, and the constant efforts by which everyone attempts to procure it, make the taste for the useful predominate over the love of the beautiful in the heart of man. Democratic nations, amongst which all these things exist, will therefore cultivate the arts which serve to render life easy, in preference to those whose object is to adorn it. They will habitually prefer the useful to the beautiful, and they will require that the beautiful should be useful."
--Alexis de Tocqueville, "Democracy in America", Part II, Chapter XI


I would have killed for a CGI Transformers movie in the 1980s.

Ordinary salary and wages are deductible for the employer and are taxed as income for individuals. Health benefits are deductible for the employer and are NOT taxed as income for individuals. It should be no mystery why we have gone through a period in which marginal health benefits have been preferred to marginal cash income.

I'm not sure this makes real-world sense. Certainly 40-60 living standards are growing faster than this graph would indicate, which basically means CPI is too high.

I think one also has to ask -- what has happened to transfer payments over this period? That is not equivalent to money to derived from work, which is much more painful to obtain, so the increase in leisure should be taken into account as well.

And I'll just point out again that the "median income" measure may be less and less meaningful as the productivity/income distribution widens -- which isn't a bad thing. Do we really want to be a country that pays ditch-diggers more rather than a country that makes billionaires out of people like Jobs and Brin?

Do we really want to be a country that pays ditch-diggers more rather than a country that makes billionaires out of people like Jobs and Brin?

I don't know. Do you really want the Third World model of a cloistered elite with kidnap insurance and helicopters to avoid all the restive, desperately poor ditch-diggers?

Why shouldn't we be a country that pays ditch-diggers more? We buy them welfare already, so let's just eliminate the rent-seeking and externalities and pay them a decent wage.

>>>so let’s just eliminate the rent-seeking and externalities and pay them a decent wage.<<<

So what exactly is your prescription? Higher minimum wages?

Our ditch-diggers are already richer than 95% of India. I don't think we're at risk of underpaying them.

Anyways, Third World countries are poor primarily because their societies lack social capital. The cloistered elite steal resource windfall wealth at gunpoint, the poor dream of being able to steal like the rich.

Oh, to answer your question: because their efforts aren't worth more, so no one would voluntarily pay them more, and forcing people to pay them more just makes everyone poorer.

Instead of ditch-diggers, we should employ diesel-powered backhoe operators, mechanical engineers, petroleum engineers, etc.

Instead of ditch-diggers, we should employ diesel-powered backhoe operators, mechanical engineers, petroleum engineers, etc.

We ultimately agree, I had more in mind the deliberate efforts to drive down American inputs per backhoe operator to the global ditch-digger mean.

Fair enough, I just think maybe we're running into a situation where the productivity gains no longer reach the semi-skilled.

I think this problem is also likely to get a lot worse as AI matures.

Ditch diggers are not at the median.

The mean becomes less meaningful as inequality widens, but the median remains as meaningful as it ever was.

"Certainly 40-60 living standards are growing faster than this graph would indicate" - and your evidence for this is?

But there was a time when ditch diggers were at the median, and if they still were, we'd be a much poorer country.

The median becomes less meaningful as productivity gains increasingly miss the semi-skilled. The mean actually becomes more important, because the rich are now subsidizing the poor to a larger and larger degree as the mean rises faster than the median.

Living standards for the 40-60 crowd are higher by virtually every measure of product access and quality -- generally >95% of the bottom half of 2012 Americans have appliances and products that were markedly inferior and just becoming common to the middle class in the 1970s, or did not even exist then.

See, for instance (and note, this isn't even the 40-60 percentiles, this is the poor):

80 percent of poor households have air conditioning. In 1970, only 36 percent of the entire U.S. population enjoyed air conditioning.
92 percent of poor households have a microwave.
Nearly three-fourths have a car or truck, and 31 percent have two or more cars or trucks.
Nearly two-thirds have cable or satellite TV.
Two-thirds have at least one DVD player, and 70 percent have a VCR.
Half have a personal computer, and one in seven have two or more computers.
More than half of poor families with children have a video game system, such as an Xbox or PlayStation.
43 percent have Internet access.
One-third have a wide-screen plasma or LCD TV.
One-fourth have a digital video recorder system, such as a TiVo.

People have forgotten how much the 1970s sucked.

How do they measure defined benefit pensions? Do they accrue value during the worker's lifetime or do they simply count them as income when they're paid? How does that compare with measurement of employer contributions to defined contribution retirement plans? Do they use the actuarial value when accrued or do they compute the actual value paid over a lifetime (which still may not be fully known).

It wouldn't surprise me to see a drop in benefits since the 70's calculated on an accrual basis because DB pensions turned out to be much more valuable than anyone expected due to longer lifespans, and they're generally no longer offered in the private sector. I wouldn't be surprised if more than half of the DB pensions promised in the 70's are currently being paid by the PBGC. But I strongly suspect the only available data is on a cash flow basis as benefits are paid out after retirement.

But I have some fundamental questions- middle class is defined in this study as the middle 20%. What happens if you look at broader measures of middle class? How has the composition of the middle 20% changed over time? Are they getting younger, are they working fewer hours, etc? Do more retirees populating the lower quintiles affect the data? What happens if you remove capital gains and dividends from income? It seems like it would be far simpler and more meaningful to look at overall wages/salaries + benefits over time. And even then it's not very meaningful because as I've said all along, inflation adjustments over long periods of time are meaningless. People seem to intuitively grasp this concept, but fancy pants economists insist on ignoring it.

1.) Why did he start the red lines (3rd Quartile family mean and mean family incomes) in 1979 but the other series in 1950? Since the red lines are indexed on 1979, it makes it look like there was no divergence between the two until 1979 and then suddenly a big change, if all you looked at was the graph you would think that median income and mean income were the same until 1979, which is wrong. In fact, that's exactly what the black lines are showing.
1b.) What's the scale on the vertical axis? 1, 2, 3?

2.) Doesn't the divergence in the black lines have something to do with the increase in household formation, especially single mothers, pulling down household income relative to GDP per capita? I guess he does mention this in his article.

3.) It's surprising how close income is to total compensation, especially considering the decade or so of double digit healthcare insurance inflation,. In fact, the solid red line doesn't seem to show any trend at all of growing over the black line, which is what I would expect to see if they were measuring the same people but one captured more income than the other. Indeed, like someone else mentioned, how could income + in kind compensation _ever_ be lower than income alone? That should set off warning bells.

4.) Shouldn't this trend also be reflected in labor's share of income? But up through 2004 at least this doesn't seem to be the case:

At the same point that the 1979 divergence of median income happens, the hourly wages of blue-collar workers also seem to stagnate.

Not sure of the reliability of this blue-collar hourly wage data but to me it looks like the stagnation has essentially got to do with the fact that the American middle class is essentially blue-collar, and globalization and resultant competition picked up around the late 70's.

I suggested in the comments at LK's site that a lot of what is going on behind that chart is that the bottom quintiles are being filled out at an faster rate than the upper ones because of the rate of increase in nonworking seniors and people on disability. As they grow, they cause the quintile parameters and the median to be consistently re-set lower, ceteris paribus. So in a sense part of this gap is just the welfare state working.

Oooh, good point. This should really be looked at through the lens of demographic shifts.

Hello,moderator.You must be a good man.I like your article
Oakley Sunglasses Clearance
okaley sunglasses

Comments for this post are closed